Intelligent infrastructure is the next innovation in delivering optimized systems for applications.
The 33rd IT Press Tour met with executives from DDN, which has built out a portfolio of infrastructure storage solutions for the enterprise -- Tintri, Intelliflash, and Nexenta.
According to Gartner, intelligent infrastructure optimizes infrastructure resources for application consumption through the use of infrastructure machine learning (ML) and applying tuning as software overlays.
By 2022, intelligent infrastructure will add infrastructure ML, analytics, and AI for IT operations (AIOps) as software overlays on top of hardware-based composable integrated systems.
Intelligent infrastructure is a new innovation in delivering optimized systems for applications.
In intelligent infrastructure, the SDI "control plane" is enhanced with automation driven by infrastructure ML to become an "intelligent plane."
AIOps:
Ingests data from multiple sources
Enables data analytics using ML at two points:
Real-time analysis at the point of ingestion (streaming analytics)
Historical analysis of stored data
Stores and provides access to the data
Suggests prescriptive responses to analysis
Initiates an action or next step based on the prescription (the result of the analysis)
According to Julia Palmer, V.P. of Research, Gartner, "the need to accommodate new workloads, integrate edge and public cloud infrastructure, and reduce operational complexity is driving I&O leaders to look for more agile integrated systems. Next-generation systems emerging today are flexible, AI-enabled, software-driven solutions that meet the requirements of digital business."
In theory, AIOps is easy as log data is analyzed with artificial intelligence to provide insights. In reality, it's complicated. ML models are immature. Insights are domain-centric. Logical unit numbers do not correlate in a virtualized environment. This results in a low and slow ROI.
Use Cases
The Institute of Cancer Research (ICR) is a world leader in identifying cancer genes, discovering cancer drugs, and developing precision radiology. They support massive amounts of data from scientific instruments and next-gen sequencers using DDN solutions.
Their challenge was to deploy a single, centralized data storage infrastructure that would enable users to analyze all types of active research with a variety of data flows. The solution is a stretched cluster with synchronous mirroring between sites. The sites are connected with private 10 Gbps network links with an option to increase bandwidth as needed. The collected data is easily accessible for analysis and big data modeling.
As the producer of high definition Major League Baseball content, MLB Network is recognized as a leader in sports video asset management. MLB Network captures and catalogs multiple petabytes of digital information each year -- 11 hours of content per game and 7,000 hours of new content each week. DDN's MEDIAScaler storage solution provides concurrent access to content while accelerating end-to-end digital content management in a single solution that can scale up to hundreds of petabytes.